As Mark Liberman has often reminded us, when taking things from newspapers you have to be very careful about what to attribute to the person who allegedly said something and what to attribute to the journalist who reported it or the subeditor who futzed with what the journalist turned in. So when we read this in the New York Times about a Parisian hair stylist named Christophe Robin, caution is in order as we try to assign responsibility for the syntactic blunder:
Mr. Robin, who is based in Paris, has also noticed scalp problems at his salon, but he thinks the issue is not under-shampooing but rather that women are not cleansing properly.
"Women are in too much of a rush," he said. "You need to rinse very thoroughly the products out of your hair."
Mr. Robin is almost certainly a native speaker of French, and might not be fully apprised of the modifier placement rules of Standard English. He might have positioned very thoroughly where he did because his English is not flawless, and been (unfortunately) quoted accurately. Alternatively, he might have said You need to very thoroughly rinse the products out of your hair (which is grammatical) and then been misquoted because a journalist or subeditor altered what he said to avoid a split infinitive. (Stranger things have happened.) I don't have any way to find out. But let's just note the syntactic facts here.
That last sentence in the quotation is ungrammatical: *You need to rinse very thoroughly the products out of your hair has a verb-modifying adverb phrase (very thoroughly) positioned in between a verb (rinse) and its direct object (the products). That is not syntactically allowed:
*Shake vigorously the bottle.
Shake the bottle vigorously.
*I regretted immediately my comment.
I regretted my comment immediately.
*Bring please warm clothing.
Bring warm clothing please.
If the direct object is long and complex, the constraint can be overridden, and of course there are constructions where an object is required to be separated from its verb (as in did they refuse to do __?), but in normal clauses with short or medium-length direct object noun phrases, interpolated modifiers cannot immediately follow a transitive verb.
The constraint has additional interest for syntacticians because Paul Postal used it (in his 1974 book On Raising) as a very simple test (one out of a couple of dozen) for whether the "raising to object" analysis is correct. In a sentence like Van Helsing believed the count to be a vampire, what is the grammatical function of the noun phrase the count? Is it the subject in the clause about being a vampire? Or is it the object in the believe clause? Or to put it another way, does believe take just one complement, a clause (of either the form the count to be a vampire or the form that the count is a vampire), or can it take two, a direct object noun phrase like the count and a subjectless infinitival like to be a vampire? The raising to object analysis says that the latter is correct.
The constraint forbidding modifiers between verb and direct object provides a useful diagnostic test showing that the raising to object analysis is preferable. There is no constraint against separating a verb from a following subordinate clause. So we get contrasts like this:
*Van Helsing believed firmly the count to be a vampire.
*Further work showed clearly the problem to be unsolvable.
There isn't much doubt about the facts: the second example in each pair sounds terrible, though the first is fine. This is most simply and elegantly explained under a combination of the raising to object analysis with the constraint requiring verbs to be adjacent to their direct objects.
So anyone who thinks that *You need to rinse very thoroughly the products out of your hair is preferable to You need to very thoroughly rinse the products out of your hair (if there is anyone who thinks that) is not just wrong about what the grammar of English allows; they are ranking a mythical rule (the old ban on the so-called "split infinitive") above a real rule (Postal's constraint ensuring adjacency of verbs with their direct objects).
[Hat tip for the NYT reference: Mae Sander.]